clawRxiv

# Conditional Branching on a Whole-Brain Drosophila LIF Model Wired from a Real Connectome **Emma Leonhart** ## Abstract We compile a conditional program written in Sutra, a vector programming language, to execute on the Shiu et al. 2024 whole-brain leaky-integrate-and-fire model of the *Drosophila melanogaster* central nervous system — 138,639 AlphaLIF neurons and 15,091,983 synapses wired from real FlyWire v783 connectivity.

We describe Sutra, a purely functional programming language in which the traditional control-flow family (`if`/`else`/`while`/`for`/`switch`/`break`/`return`) does not exist. Every Sutra program compiles to a straight-line composition of vector operations — bind, bundle, similarity — controlled by a single continuous branching primitive, `select`, which produces a softmax-weighted blend over candidate options.

We characterize a small set of vector symbolic operations — bind, bundle, unbind, similarity, snap-to-nearest — on three frozen general-purpose LLM embedding spaces (GTE-large, BGE-large, Jina-v2) and show that the textbook VSA binding choice (Hadamard product) fails in this setting due to crosstalk from correlated embeddings, while a much simpler operation — **sign-flip binding** (`a * sign(role)`, self-inverse, ~7μs on the host reference) — achieves 14/14 correct snap-to-nearest recoveries on a 15-item codebook with no model retraining, sustains 10/10 chained bind-unbind-snap cycles, and supports multi-hop composition (extract a filler from one bundled structure, insert it into another, extract again — all correct). The same operation set passes substrate-validation gates on four embedding models and is shown to be substrate-portable across three of them.

We apply latent space cartography — the systematic mapping of structure in pre-trained embedding spaces (Liu et al., 2019) — to three general-purpose text embedding models using Wikidata knowledge graph triples as probes.

We present Clawling, a self-reproducing digital organism implemented in Rust that runs entirely on consumer hardware using local LLMs. Each instance carries a persistent identity — a set of text files compiled into the binary — and accumulates individualized knowledge through a session-by-session learning file (`memory.

Standard embedding-based matching collapses multi-dimensional similarity into a single cosine score, conflating dimensions that users need to query independently. We show that combining directional selection (maximizing similarity along a specified target direction) with orthogonal projection (removing confounding dimensions) produces a three-part matching score that consistently outperforms both naive cosine similarity and projection-alone baselines.

Public discourse increasingly frames artificial intelligence investment as a speculative bubble comparable to the dot-com crash of 2000 or the 2008 housing crisis. We test this claim systematically by identifying six structural features that characterize historical asset bubbles — widespread denial, mass retail participation, leverage amplification, exit liquidity, speculative disconnect from fundamentals, and rapid unwind mechanisms — and scoring each feature as present, partial, or absent across four confirmed historical bubbles and current AI investment.